S1995
Physics - Dose prediction/calculation, optimisation and applications for photon and electron planning
ESTRO 2026
Proffered Paper 4363 Optimizing dose to blood: automated planning to spare blood-bearing organs associated with survival in locally advanced non-small cell lung cancer Shu Xing 1 , Masoud Zarepisheh 1 , Gourav Jhanwar 1 , Sharif Elguindi 1 , Narek Shaverdian 2 , Joseph Deasy 1 , Maria Thor 1 1 Medical Physics, Memorial Sloan Kettering Cancer Center, New York, USA. 2 Radiation Oncology, Memorial Sloan Kettering Cancer Center, New York, USA Purpose/Objective: This study aims to identify blood-bearing organs whose mean blood dose significantly associated with overall survival (OS), and to evaluate the feasibility of sparing these organs during treatment planning. Material/Methods: A total of 155 patients with locally advanced non-small cell lung cancer (LA-NSCLC) treated with concurrent chemoradiotherapy (chemo-RT) were retrospectively analyzed. The median prescription dose was 63 Gy, delivered in 1.8-2 Gy fractions with 5-field intensity- modulated radiotherapy (IMRT). An in-house Artificial Intelligence (AI) auto-segmentation algorithm was used to delineate 19 blood-bearing structures (Figure 1). Hematological Dose (HEDOS) computational framework [1,2] was used to estimate blood dose for each structure. The top seven major contributing organs to blood dose were identified, for which Cox proportional hazards regression was performed to assess their association to OS (significance: p- value<0.05). Mean blood dose to lung, heart, aorta and pulmonary artery (PA) demonstrated significant correlation with OS. A 10-patient replanning study (5 male and 5 female) was subsequently conducted to evaluate the feasibility of sparing these organs. All 10 LA-NSCLC patients received concurrent chemo-IMRT, with 60 Gy in 2 Gy fractions. Both the original clinical plan and the re-optimized plan were generated using our in-house automated planning system -- Expedited Constrained Hierarchical Optimization (ECHO) [3], ensuring unbiased comparisons. With the lung and heart dose constraints following strict clinical guidelines, additional optimization objectives were introduced for the aorta and PA. The resulting differences in target coverage and organ-at-risk dose were evaluated.
Results: The median OS for the 155 patient cohort was 1.8 years with interquartile range (IQR) of 2.9 years. Increased mean blood dose to the lung, heart, aorta and PA significantly predicted worse OS (Hazard ratios=1.1-5.3; p-values=0.003-0.02). Dose constraints for aorta and PA are not part of standard clinical practice. Using the population mean doses (48Gy for aorta, 56 Gy for PA) as new optimization constraints, mean doses to the aorta and PA were reduced by a median (IQR) of 3.0 (2.3) Gy and 3.6 (4.3) Gy, respectively (Figure 2). Target coverage was maintained with a negligible median difference of 0.08 Gy, and all clinical dose constraints were met.
Made with FlippingBook - Share PDF online